Auto-launching of applications

[Jordan Mantha]

On Wed, Feb 25, 2009 at 1:43 AM, Mark Shuttleworth <mark at ubuntu.com> wrote:
> Oliver Grawert wrote:
>
> so why dont you wait until you got proper hard data and come up with a
> real revolutionary new design for the problem based on this data and
> instead switch to an half thought through immature, resource hogging
> and user annoying mechanism, sorry if that sounds harch but for me that
> new beahavior is a big step backwards in usability, software needs to be
> out of the way of the user, i dont want my hammer to discuss with me
> about the nail, i just want to use it to get the job done ...
>
>
> Oliver, this was not constructive feedback. If you think your email sounds
> harsh, please sleep on it and reconsider it in the morning. The work we are
> doing fundamentally depends on making changes to existing patterns. That
> will be difficult, please don't make it more difficult by raising the blood
> pressure of the discussion without providing clear suggestions for ways to
> improve things.

While it might have been a bit stronger tone than you'd like I do
think there is definitely constructive feedback in what Oliver said.
That we are having this discussion this far into Jaunty's development
cycle should be saying something. While I'm very appreciative of your
emails today explaining a bit more about what's going on, a "opps,
pardon the mess" notice this late in the release after it's too late
to actually change course is not exactly reassuring.

I think what Oliver was saying was that it might have been wiser to
spend Jaunty nailing down the notification vision and implementation
and communicating that to the developer community for feedback, then
landing code earlyish in Karmic so people can play around with it.
What I mostly saw communicated to the development community was "trust
us, you're gonna love it!". My general experience is that development
communities don't like surprises, especially half-baked ones, for core
desktop experience systems.

So, how can we make sure that the development community at-large is
adequately notified, informed, and engaged about these sorts of
changes with sufficient time for feedback? I've seen quite a few "huh,
we discussed it at UDS, I don't know what you're complaining about"
type statements on IRC concerning various specs. I'm wondering if the
"results" of UDS are not being adequately communicated to the rest of
the development (or say contributor) community. Also, as specs are
"tweaked" after UDS how does that information get disseminated? Should
we perhaps have, for instance, an RSS feed for spec changes (on
approved specs anyway)? I would have surely enjoyed a set of
"Notification Roadmap" blog posts with some rough milestones. There
were some mockups done and that's good, but I think timing was missing
for me at least. A set of "known issues but we're gonna land anyway"
emails may have helped. Any other thoughts?

-Jordan